Initializing list of numpy arrays

Question

I am trying to initialize a list of numpy arrays in the following way

import numpy as np
sol=[np.zeros(5)]*4

But when I try to modify one of the list members like this,

sol[0][2:4]=[1,1]

It changes all the list members instead of only the first one

[array([ 0.,  0.,  1.,  1.,  0.]), array([ 0.,  0.,  1.,  1.,  0.]), array([ 0.,  0.,  1.,  1.,  0.]), array([ 0.,  0.,  1.,  1.,  0.])]

I guess it is old story about mutable and inmutable objects, but I simply don't know how to solve it.

I tried using copy and deepcopy, but not successful.

Sorry if it is duplicated entry, but I couldn't find any similar question

Thanks

You need to create *a separate `array`* each time: `sol = [np.zeros(5) for _ in range(4)]` — jonrsharpe, Jul 07 '14 at 16:04

score 2 · Accepted Answer · answered Jul 07 '14 at 16:04

2

and the answer is:

np.zeros((4, 5))

and to explain the point about mutable objects. when you do this:

[np.zeros(5)] * 4

it's functionally the equivalent of this:

a = np.zeros(5)
[a, a, a, a]

answered Jul 07 '14 at 16:04

acushner

1

Dammit!! i guess I am tired. Thanks for the mutable explanation. – manolius Jul 07 '14 at 16:12
2

remember you can always check an object's id with function `id`. if those numbers are the same, then it's the same object. – acushner Jul 07 '14 at 16:20

1 Answers1